Deciphering molecular heterogeneity and relevance of subtypes in breast cancer progression

Breast cancer affects over 3500 women in Norway every year. Since the introduction of mammographic screening, early stage breast cancers are diagnosed more frequently and many of these are non-invasive with low potential of spreading and affecting mortality. Since there are currently no reliable biomarkers that predict the ability of a non-invasive cancer to invade, many patients are likely overtreated. In this thesis, Helga Bergholtz and colleagues have performed molecular studies of different stages of breast cancer and characterized the heterogeneity of tumor progression. They have analyzed normal breast tissue biopsies from two time-points and studied the transcriptomic changes over time in relation to mammographic density. They have also analyzed gene expression, DNA copy number aberrations and mutations in a chemically induced mammary gland tumor model in mice and show that this may be a suitable model for studying a special type of breast cancer called claudin-low. The thesis has special emphasis on a non-invasive cancer called ductal carcinoma in situ (DCIS). They have performed extensive molecular characterization of DCIS compared to invasive breast tumors using gene expression, DNA copy number and DNA methylation data. Since molecular subtypes have strong clinical relevance in breast cancer, the analyses were performed stratified by subtype. This revealed that basal-like DCIS and basal-like invasive tumors were distinctly different, and that basal-like DCIS might not be precursors of basal-like invasive tumors. Also, they have performed targeted sequencing of different cellular compartments in mixed DCIS/invasive tumors and show a similar mutation spectrum in the two tumor compartments. The findings in this thesis may impact the understanding of tumor progression and contribute to more precise therapy for patients with DCIS